Arrival Time from Spring Migration in Male Pied Flycatchers: Individual ...

The Condor 100:702-708 D The Cooper Ornithological

Society 1998

ARRIVAL TIME FROM SPRING MIGRATION FLYCATCHERS: INDIVIDUAL CONSISTENCY RESEMBLANCE’

IN MALE PIED AND FAMILIAL

JAIME POT-IT Departamentode Biologia Animal, Universidadde Alcalri, E-28871 Alcalci de Henares (Madrid), Spain, e-mail: [email protected] Abstract. The timing of arrival from spring migration was studied in 4 years for male Pied Flycatchers (Ficedula hypoleuca)in central Spain. Heritability and repeatability analyses were performed at several levels of familial resemblanceand male age in order to determine whether arrival times are consistent within individuals between years and are similar between relatives. Oldest males arrived earlier. Arrival time explained much of the variation in laying time and number of young fledged. Age-independent variation in wing length also affected arrival date, the males with longest wings settling earlier. Arrival times were not repeatablewithin individuals acrossyears and were not similar between relatives (parents-sons,full-sibs). Although genetic variation may exist in departure dates of longdistancemigrant birds, the close connectionof an early arrival to high reproductivesuccess may have depleted genetic variation in arrival time. In addition, environmental variation probably is too high to detect significant heritability in arrival times without very large sample sizes. Key words: age, Ficedula hypoleuca, heritability, migration, Pied Flycatcher, repeatability.

INTRODUCTION Each year migratory birds must find, occupy, and defend a breeding territory on their arrival from winter quarters. Among passerines,a common pattern is for males to arrive earlier than females, most likely due to within-sex competition for accessto resourcesvital for reproduction, and older birds arrive earlier than younger ones (Francis and Cooke 1986, Hill 1989, Lozano et al. 1996). Early arrival at the breeding sites increasesreproductive fitness through more nest sites (Potti and Montalvo 1991a) or better resources(Slagsvold 1986, Aebischer et al. 1996), earlier pairing (Slagsvold and Lifjeld 1988, Potti and Montalvo 1991a) and, in promiscuous and/ or facultatively polygynous species, greater opportunity to attract additional females (Alatalo et al. 1981, Potti and Montalvo 1993). Early arrival enhancesmale and female breeding success which is accentuatedby a widespread tendency for early breeders to have a greater chance to fledge and/or recruit more offspring (Hochachka 1990, Meller 1994, Aebischer et al. 1996). Although much research has been conducted on intrinsic and extrinsic factors affecting bird migration, the evolutionary important question ‘Received 1998.

18 August 1997. Accepted 15 April

whether components of migratory behavior in animals are heritable, and thus capable to respond to selection, has only recently come under study (Berthold 1991, Dingle 1994). Experimental work with captive European Blackcaps (Sylvia atricapilla) has shown that some components of the migratory syndrome, such as migratory restlessness(Zugunruhe) or migratory direction, are inherited and are probably polygenie traits with threshold effects (Berthold 1991, Pulido et al. 1996). Data on individual (ontogenetic) and familial variation in migratory behavior in wild bird populations are scant (Rees 1989) and hard to obtain, due to high dispersal,low site fidelity, short life spans, and long-term work required. Here, I report data on one component of migratory behavior, the timing of arrival of males on the breeding grounds, that were obtained acrossfour years for a population of Pied Flycatchers (Ficedzda hypoleuca). Pied Flycatchers are small (- 12 g) long-distance migrants overwintering south of the Sahara and breeding across the western Palearctic (Lundberg and Alatalo 1992). I investigate whether there exist individual consistency or stability (repeatability) and familial resemblance (heritability) of male arrival time. I found no clear evidence that arrival date of males is a heritable trait in this species.

17021

HERITABILITY

METHODS I studied the arrival times of male Pied Flycatchers through the spring and early summer from 1988-1991 in a population breeding in nestboxes in a mature deciduous (Quercus pyrenaicu) oakwood in central Spain (Potti and Montalvo 1991a). The area was censuseddaily from mid April onwards to document exact arrival dates and identity of males. Soon after arrival, males start to sing at one or more nestboxes, making census of males a relatively easy task. Since 1987 most arriving males were previously colorbanded, allowing their quick identification. The majority of remaining males only wore a metal band or were unbanded and were distinguished until capture by age-distinctive plumage traits (Potti and Montalvo 1991a), by idiosyncratic features of their plumage color (Potti and Merino 1993, by the size and form of the white patch on the forehead (Potti and Montalvo 1991a), or by the position of a metal band put on them as a fledgling in previous years (since 1984). It is very rare for a Pied Flycatcher to lose its preferred nestbox to a conspecific or interspecific competitor and, when this occurred, the same male was usually observed again singing in a neighbor nestbox. Hence, it is likely that few errors were made in the a posteriori assignment of arrival times to previously unbanded birds. Males were captured inside their nestboxes while feeding nestlings. All birds were individually marked with color bands and aged as yearlings or older (Karlsson et al. 1986, Potti and Montalvo 1991a), and their wing length (flattened chord; Svensson 1992) was measured.Immigrant birds aged as “older” at first capture were assigneda minimum age of two years (Potti and Montalvo 1991b). The laying date of a particular nest was defined as the day the first egg in the clutch was laid. Laying dates and number of young fledged varied across years and were standardized by using the deviations of individual values from median annual values (laying date) or the residuals (deviations from calculated group averages) of an analysis-ofvariance (ANOVA) in relation to study year (young fledged). To calculate repeatabilities (see below) of laying dates in nests of individual males in different years, I used the entire data set from my long-term population study, comprising breeding records from 1987 to 1995.

OF ARRIVAL

DATE

703

Arrival times were standardized within years by subtracting data points from median annual values and then transforming them to z-scores with 0 mean and unit variance (hereafter standardized arrival time). I first estimated familial resemblance in arrival times by calculating the repeatability. Repeatability is estimated by making repeated (annual, in this case) measuresof a sample of individuals and then calculating the ratio of the among-individual variance to the sum of both the among-individual and withinindividual variances (the intraclass correlation coefficient; Zar 1996). Components of variance are estimated by one-way ANOVA. A high repeatability indicates that variation within individuals is much smaller than among individuals. If measurementswithin individuals are very different, repeatability is low. I used linear regression to compute heritability (W) of the arrival times of sons on those of their presumptive fathers. Twice the slope of the regression estimates the heritability (Falconer 1986); the standard error (SE) of the regression slope also was doubled. Arrival times of full sibs also were compared using one-way ANOVA, as in the analysis of repeatability. Heritability was computed as twice the intraclass correlation coefficient (Falconer 1986). In the analyses of parent-son resemblance, when two or more data points for a particular parent or family (sibships) were available, I used averages of standardized data to avoid pseudoreplication. To estimate the power (Thomas and Juanes 1996, Zar 1996) of heritability estimates,that is, the probability of correctly rejecting the null hypothesison absenceof heritability given a specified significance level (alpha = 0.05) and sample sizes, as well as the sample sizes needed to detect her&abilities of specified magnitude with a given power, I used program PS of DuPont and Plummer (1990) for linear regression (available as freewareon the Internet; ftp://ftp.vanderbilt.edu/ pub/biostat/ps.zip). The power of analyses of repeatability was estimated by means of the Pearson correlation coefficient, using only males with arrival dates known in two years. Statistical tests are two-tailed. RESULTS The first males arrived on the study area in the third week of April; successfully pairing males can arrive as late as 12 June (Potti and Montalvo 1991a). Arrival times of males gradually ad-

704

JAIME POTTI

42

68 1

45

I

Male age (years) FIGURE 1. Arrival times (means? SE) of male Pied Flycatchersin relationto age. Date 30 is April 30. Numbersabovelines are samplesizes. ’

vanced with increasing age (Fig. I), although 0.22, n = 31 and 133 birds, respectively), nor significant differences were apparent only be- between presumptive fathers and sonswithin the tween yearlings and older birds (Tukey a pos- subsample of familial data (P = 0.21, n = 22 teriori tests). Standardized arrival times were families). I obtained repeated records of arrival time for earlier for longer-winged males (I = -0.30, IZ = 164, P < O.OOl), which might be confounded 39 males, 32 of which had 2 recordsin different, by age due to the significantly longer wings of usually successiveyears, 4 males had 3 annual older birds (Alatalo et al. 1984a). However, ar- records, and 3 males had 4 records. The repeatrival time was earliest for males with longer ability (R) of arrival times was very low and not wings even after controlling for the effect of age significant (R = 0.04; mean number of data per (partial correlation, I = -0.21, P = 0.006). This individual, no = 2.25, F38.87 = 1.10, P > 0.30). indicates that the effects of increasing age and The same was true when standardized (R = wing size on an earlier arrival date were at least 0.08; Fx.8, = 1.19, P > 0.25) and residual ar= 1.21, P = 0.27; rival times (R = 0.08; F38,87 partially independent. Being early was advantageousfor male fitness power > 0.10) were used. In order to remove as the number of young fledged was lower for any potential confounding effects of variation in males that arrived later in the season(r = -0.38, age and arrival time among males, I selected,for n = 138, P < 0.001, Fig. 2). There was no cost further confirmatory analysis, repeated records in being too early, as indicated by the nonsig- of arrival only for males 2 or more years old. nificant quadratic regressionterm (quadratic co- The repeatability of residual arrival time in this data set was again nonsignificant (R = 0.09; no efficient = -0.002, P > 0.10; Fig. 2). In order to remove variation in arrival times = 2.18, F26,58 = 1.22, P > 0.25), indicating that due to age, I further standardized arrival times individual adult males are not consistent in the for familial comparisons by using the residuals relative timing of their arrivals on the breeding of an ANOVA of standardized arrival time (see grounds as they age. Methods) in relation to male age (hereafter reThere were 32 male sons in 22 families for sidual arrival times). Residual scores of arrival which one or more arrival times of the sons and time were normally distributed (Kolmogorov- their presumptive male parents were known. AlSmimov tests, P > 0.25) and their distribution though the relationship was positive, the heritadid not differ between young and older birds bility of residual arrival times did not differ sig(Kolmogorov-Smirnov two-sample tests, P = nificantly from zero (h* = 0.34 + 0.73, P >

HERITABILITY

OF ARRIVAL DATE

705

3. .

x2 B a, lC= p CO 5

. .

p. . . . . . . y.. . . . . S.. . . . . .. .

.

-1

I.

:

. .

p -2 -

.

$j

.

.

.

.

2 -3 Z m 5 -4 -

.

-20

. .

-10

.

.

.

.

.

.

.

0

10

Standardized

arrival date

30

20

FIGURE! 2. Relation between standardizeddate of arrival of male Pied Flycatchersand standardizednumber of young fledged (see Methods for standardizedcomputations).The line is the least-squaresregressionline (regressionfunction: y = 0.11-0.05x).

0.60, n = 22, Fig. 3). However, given the large variances of arrival dates, the power of this test at the customary alpha level of 0.05 is very low (power = 0.07). Furthermore, that value of h* is entirely due to the “pull” of an outlier (Fig. 3); when this point is excluded a “no sense,” negative heritability is obtained (h* = -0.50 + 1.00,

There were 9 families for which the arrival times of 2 or 3 (presumptive) full-sib males were known. The heritability of residual arrival dates calculated on the basis of (presumptive) full-sib resemblance was h* = 0.08 (no = 2.25, FS,,9 = 1.09, P > 0.40). As there were more data available on laying dates, which are closely related to male arrival

P > 0.60, R = 21).

I

01 I

3 zm 75 .! z 2 5

ff = 0.34,SE

2-

= 0.73,P = 0.64

0 0

I I

01 I I I I I I 0 I I I

l0

0

-

0* ’ -------------__i___________

! n 2 oz

0

-1

’

0

0

q:

I

I

, I

-2

-1

0

0 0

/ 1

Residual arrival time of presumptive father

FIGURE 3. Relationship between standardizedarrival times of male Pied Flycatchers and their presumptive sons (see Methods for standardizedcomputations). The line is the least-squaresregression line (regression function: y = 0.29 + 0.17x).

706

JAIME

POTTI

netic parentage are available for my population and, unfortunately, the method to compute EPF rates on the basis of differences in heritability estimates of morphological traits using motherand father-on-offspring regressions (Alatalo et al. 1984a) has proven unreliable to that end (Hasselquist et al. 1994, Potti and Merino 1994). However, misassigned parentage is not a problem for repeatability estimates, which indicated an overall inconstancy of male arrival and laying dates, and hence a likely absence of detectable genetic variation (Boake 1989). Checking for the existence of repeatability is a first step in proposed protocols to detect genetic variation because repeatability sets an upper limit to the heritability of a trait, which cannot be larger than the former (Falconer 1986, Boag and van Noordwijk 1987). In addition, because genetic variation within individuals is absent, EPFs are not a confounding factor in the computations of repeatability, although they certainly are a problem for the quantification of heritability. In spite of this, the consistency of nonsignificant repeatability and heritability estimatesin this study indicates that, whatever may be the rate of EPFs DISCUSSION in my population, genetic variation for arrival I have presented what to my knowledge are the time, if present, must be low. first data obtained on familial resemblance in There are, however, further reasonsto be cautiming of arrival to the breeding sites in a long- tious about the absenceof genetic variation, bedistance migrant passerine. The estimates were cause it may exist but be masked by genotypenot confounded by the social environment (Rees environment interactions, which are likely to be 1989) because Pied Flycatchers live solitary presentin such a complex “trait” as arrival date. lives after they become independent of their par- Arrival times must depend to some degree on ents, and they migrate alone (Lundberg and Aladistances migrated and departure dates from talo 1992). Using a large data set, Rees (1989) winter quarters, which in turn depend upon enshowed that arrival and departure times of winvironmental variables, including photoperiod tering social Bewick’s Swans (Cygnus columbiand weather conditions there. In addition, the U~US [bewickii]) had a slight heritability (h* = predictably high environmental variation met by 0.10-o. 19), although in this case the estimate the birds while on passage in their migratory could to some degree be confounded by the so- journey, and the likely important roles played in cial habits of swans and their tendency to mi- migration performance by bird age and previous grate in groups,hence cultural, not genetic trans- experience with both the migration route and the mission. This problem did not exist in my study, area (Montalvo and Potti 1992, Cantos and Teland I still was unable to find evidence for ge- lerfa 1994), may sum up to a high stochastic netic variation in the arrival times from spring variance in timing of arrival (Aebischer et al. migration in male Pied Flycatchers. 1996). Body condition (Arvidsson and Neergard One may argue that lack of father-son resem- 1991), including lack of infection by hematoblance in arrival dates may be due to genetic zoan parasites (Ratti et al. 1993), and body size analyses being plagued by a high frequency of (wing and first primary lengths; Alatalo et al extra-pair fertilizations (EPF), which are known 1984b, Potti and Montalvo 1991a) also are reto occur in the Pied Flycatcher at variable rates lated to early arrival in at least some bird spe(Lifjeld et al. 1991, Gelter and Tegelstriim cies. For instance, male flycatchers with the lon1992). No DNA fingerprinting estimates of ge- gest wings arrived earlier independently of their

and pairing dates (Potti and Montalvo 1991a), I investigated whether laying times for 142 males with 2-6 breeding records were repeatable. The repeatability was -0.00 (F,,,,,,, = 0.99, P > 0.50) after standardizing the data set for significant variation with respect to year and female and male ages by using residuals of a factorial ANOVA. I also correlated the average residual laying dates in nests of sons with those of their presumptive male parents. The resemblance in laying date in nests of male parents and male offspring was nil (r = -0.01, it = 74, P > 0.90; power = 0.05) and the same was true when the comparison was made with average residual laying dates of their mothers (controlling for significant variation of laying date with female age; I = -0.07, P > 0.70, IZ = 103 males in 71 families; power = 0.06). Hence, there is no evidence that arrival and laying times may be both repeatable or heritable in male Pied Flycatchers. However, the chancesto correctly reject the null hypothesis of no heritability of male arrival and laying dates were very low (power = 0.050.12).

HERITABILITY

OF ARRIVAL

DATE

707

age, which may be interpreted as evidence that that affect it plus additional environmental varia size constraint may operateon migration speed ation, such that a low heritability is to be exand counters the argument that late arrivals by pected (Price and Schluter 1991). For instance, young, small-winged birds are due to their rel- wing length seems to retain genetic variance in ative inexperience with the migratory journey. this population (around 0.50; unpubl. data), An analytical problem in this study was the whereas corresponding variation in arrival time, low power of statistical teststo detect significant one of its correlates, is almost undetectable. This heritability (Falconer 1986, Arnold 1994). Pow- illustrates that the link from genes to wing er calculations indicate that very large sample length to a life-history trait such as arrival date sizes are necessary to detect a low heritability is long: much environmental variance is added with the sample variance observed. For exam- at each step in the path of morphological to lifeple, to detect a low heritability of, say, 0.10 (re- history traits, so that even though there may be gression slope = 0.05) with a reasonably high a genetic component to arrival date via its con(Thomas and Juanes 1996) power of 0.80 (that nection with heritable body traits, this must be is, an 80% chance of getting a significant result very small (Price and Schluter 1991, Roff 1997). given that the null hypothesis of no heritability It is not possible to distinguish between the “exis false), it would be necessary to have data on haustion of genetic variance hypothesis” and the arrival dates for 641 families, implying about 32 “environmental noise hypothesis” with the data years of data collection at the current recruit- at hand and, most likely, both may be true. ment rate in my population (Potti and Montalvo Genetic variation may underlie the proximate 1991b). If heritability of arrival date was larger, control of departure dates in long-distance misay 0.50 (a common heritability for morpholog- grants, as well as in birds migrating shorter disical traits, Boag and van Noordwijk 1987), then tances (Berthold 1991), but still be undetectable 335 father-son pairs (about 17 years of field with arrival times to the breeding grounds. Inwork) would have to be sampled to detect it at dividuals may not be consistent in their time of alpha = 0.05 and power = 0.80. Similar com- arrival from one year to another because computations for detecting a low (h2 = 0.10) heri- plex interactions between environmental factors tability of “male laying dates” result in un- affecting arrival date and migration perforwieldly figures (> 13,000 and > 295,000 data mance. Demonstrating among-population genetpairs, using laying dates of mother and father ic variation in major traits, such as c&annual rhythms (Berthold 1991, Pulido et al. 1996), can Pied Flycatchers, respectively). The fact that male arrival time in spring is help to explain within- and between-species closely connectedto male fitness in the Pied Flyvariation in the urge to migrate and hence on catcher (Fig. 2) might be taken as a priori evi- departure dates to breeding or wintering areas dence in support of a predictable absenceof ge- (Ellegren 1990). However, environmental varinetic variation in this characteristic. Genetic ation will likely make it difficult to demonstrate variance might have been depleted by the close within-population genetic variation in arrival association of early breeding to fitness due to times of migrant birds in natural settings. strong and constant selection, with all remaining ACKNOWLEDGMENTS variation being enviromentally caused (“Fisher’s fundamental theorem;” Mousseau and Roff I thank the reviewers for detailedcommentsimproving 1987). The evidence presented here and back- a first draft. The study was funded by grants of the ground information on this population are con- SpanishDireccidn General de Investigaci6n Cientifica y Ttcnica (projects PB86-0006-C02-02 and PB91sistent with this interpretation. An early date of 0084-CO3-03). During preparationof the manuscriptI breeding is the most important determinant of was supportedby project PB94-0070-C02-02. recruitment in this population (Potti and Montalvo 1991b; unpubl. data), imposing a strong LITERATURE CITED selection pressure to settle and breed early. AlAEBISCHER,A., N. PERRIN, M. Kamo, _I.STULIER,AND D. R. MEYER. 1996. The role of territorychoice,mate ternatively, as discussedabove, low heritability choice and arrival dam on breedingsuccessin the of arrival date may be due to the predominant Savi’s WarblerLocustellaluscinioides. J. Avian Biol. importance of environmental variables affecting 27:143-152. it (Price and Schluter 1991), i.e., arrival times ALATALO, R.V., A. CARLSON,A. LUNDBERG, AND S.ULFsm.4m. 1981. The conflictbetweenmale polygamy may be subject to all the variation in the traits

708

JAIME

PO-IT-1

and female monogamy:the case of the Pied Flycatchersrevealed by DNA fingerprinting.Behav. catcherFicedzdahypoleuca.Am. Nat. 117:738-753. Ecol. Sociobiol.29:95-101. G. A., S. PERREAULT, ANDR. E. &ON. 1996. ALATALO, R. V., L. ~JSTAFSSON, AND A. LUNDBERG L~Z.ANO, 1984a. High freouencvof cuckoldrv in Pied and Age, arrival date and reproductivesuccessof male CollaredI&atchers. dikes 42:41-4?. American RedstartsSetouhaparuticilla. J. Avian Biol. 27:16&170. - ALATALO,R. V., L. GUSTAFSSON, AND A. LUNDBERG. A., ANDR. V. AL_ATALO. 1992. The Pied Fly198413.Why do young passerineshave shorter LUNDBERG, wings than older birds?Ibis 126:410-415. catcher.T & A. D. Poyser,London. ARNOLD, S. J. 1994. Multivariate inheritanceand evoMILLER, A. P 1994. Phenotype-dependent arrival time lution: a review of concepts,p. 17-48. In C. R. B. and its consequencesin a migratory bird. Behav. Ecol. Sociobiol.35:115-122. Boake [ed.], Quantitativegenetic studiesof behavioral evolution.Univ. ChicagoPress,Chicago,IL. MONTALVO, S., ANDJ. Porrt. 1992. Breeding dispersal hvIDSSON, B. L., ANDR. NFERGARLX 1991. Mate choice in SpanishPied FlycatchersFicedulahypoleuca.Ornis Stand. 23:491-498. in the Willow Warbler-a field experiment.Behav. MOUSSEAU, T A., ANDD. A. Row. 1987. NaturalselecEcol. Sociobiol.29:225-229. tion and the heritabilityof fitnesscomponents.HeBEXTHOLD, l? 1991. Geneticcontrolof migratorybehavredity 59:181-197. iour in birds.TrendsEcol. Evol. 6:254-257. BOAG,l? T, ANDA. J. VAN N~~RDWUK.1987. Quami- PATTI,J., AND S. MERINO. 1994. Heritability estimates and maternaleffects on tarsuslength in Pied Flytative genetics,p. 45-78. In E Cooke and F! A. catchers,Ficedula hypoleuca.Gecologia 100:33lBuckley [eds.], Avian genetics. Academic Press, London. 338. BOAKE,C. R. B. 1989. Repeatability:its role in evolu- Porn, J., ANDS. MEIUNO.1995. Some male Pied FlycatchersFicedula hypoleucuin Iberia become cdltionary studiesof mating behavior.Evol. Ecol. 3: lared with age. Ibis 137:405-409. 173-182. CANTOS, E J., ANDJ. L. TELLERS.1994. Stopoversite Porn, J., ANDS~MONTALVO.1991a. Male arrival and female mate choicein Pied FlycatchersFicedulahyfidelity of four migrantwarblersin the lberian peninsula.J. Avian Biol. 25:131-134. poleucain centralSpain.Grnis Stand. 22:45-54. DINGLE,H. 1994. Geneticanalysesof animal migration. Porn, J., ANDS. MONTALVO.1991b. Returnrate, age at p. 145-164. In C. R. B. Boake, [ed.], Quantitative first breedingand nataldispersalof Pied Flycatchers genetic studiesof behavioralevolution.Univ. ChiFicedulahypoleucain centralSpain.Ardea 79:419cago Press,Chicago. 428. DUPONT,W. D., ANDW. D. PLUMMER JR. 1990. Power Porn J., AM) S. MONTALVO.1993. Polygynyin Spanish and samplesize calculations:a review and computer Pied FlvcatchersFicedulahvooleuca. Bird Studv ,* . 40: program.ControlledClinical Trials 11:116-128. 31-37: ELLEGREN, H. 1990. Timing of autumn migration in PRICE,T, ANDD. SCHLUTER. 1991. On the low heritaBluethroatsZawinia s. svecicadependson timing of bility of life-history traits.Evolution45:853-861. breeding.ChnisFennica67:13-17. Pu~mo,E, I?BERTHOLD, ANDA. J. VANNOORDWUK. 1996. FALCONER D S. 1986. Introductiona la geneticacuanFrequencyof migrantsand migratoryactivityaregetitativa.Cecsa Mexico D.E. Mexico. neticallycorrelatedin a bird population:evolutionary FRANCIS, C. M., ANDE COOKE.1986. Differentialtiming implications. Proc. Natl. Acad. Sci. 93: 14642of springmigrationin wood warblers.Auk 103:54814647. 556. R0r1, O., R. DUFVA,ANDR. V. ALATALO.1993. Blood GELTER,H. l?, AM) H. TEGELSTR~M. 1992. High heparasitesand male fitness in the Pied Flycatcher. quency of extra-pairpaternityin Swedishpied flyOecologia96:410-414. catchersrevealed by allozyme electrophoresisand REES,E. C. 1989. Consistencyin the timing of migration DNA fingerprinting.Behav.Ecol. Sociobiol.31:1-7. for individual Bewick’s Swans. Anim. Behav. 38: GWINNER, E. 1977. Circannualrhythms in bird migra384-393. tion. Annu. Rev. Ecol. Syst. 8:381-405. ROFF,D. A. 1997. Evolutionaryquantitativegenetics. HASSELQUIST, D., S. BENSCH, ANDT VON SCHAN~Z.1994. Chaumanand Hall, New York. Estimatingcuckoldryin birds:the heritabilitymeth- SLAGSVOL~, T 1986. Nest settlementby the Pied Flyod and DNA fingerprintinggive differentresults.Oicatcher.Does the female chooseher mate for the kos 72:173-178. quality of his house or himself! Gmis Stand. 17: HILL, G. E. 1989. Late spring arrival and dull nuptial 210-220. plumage:aggressionavoidanceby yearling males? SLAGSVOID, T., ANDJ. T L~;JELD.1988. Plumagecolour Anim. Behav. 37665-673. and sexualselectionin the Pied FlycatcherFicedulu HOCHACHKA, W. 1990. Seasonaldeclinein reproductive hypoleuca.Anim. Behav. 36:395-407. performanceof Song Sparrows.Ecology 71:1279S~ENSSON, L. 1992. Identificationguideto Europeanpas1288. serines.Svensson,Stockholm. KARLSSON L., K. F%RSSON, AND G. WALIN~ER.1986. THOMAS,L., ANDE JUANES.1996. The importanceof Ageing and sexingin Pied FlycatchersFicedulahystatisticalpower analysis:an examplefrom Animal poleuca.Var Fagelvaild45:131-146. Behaviour.Anim. Behav. 52:856859. LIFJELo,J. T, T SLAGSVOLD, ANDH. M. LAMPE.1991. ZAR, J. H. 1996. Biostatisticalanalysis.3rd ed. Prentice Low frequencyof extra-pairpaternity in pied flyHall, EnglewoodCliffs, NJ.